A Critique of Horgan’s Identity for Phenomenal State-types and Neurophysiological State-types

Terence Horgan, in “Functionalism, Qualia, and the Inverted Spectrum,” prescribes a view of the mind which combines functionalism for non-phenomenal states and type-type psychophysical identity for phenomenal states.[1] I do not buy into the explanatory power offered by type-type psychological identity for phenomenal states, and thus I can’t agree with that aspect of Horgan’s program. There are two basic reasons for my opposition.  One, I am not sure how a type-type psychophysical identity view can explain phenomenal states, since I don’t see how neurophysiology is relevant to an explanation of one’s phenomenology (I’ll explain what I mean by this later); two, on a more general note, psychophysical identities usually seem to presuppose that the “Unity of Science” view is appropriate for psychological theories, and I oppose that view for several reasons.[2]  In what follows, I review the basics of Horgan’s view, and subsequently argue against the appropriation of phenomenal states according to psychophysical identity in light of the articulation and explication of the aforementioned reasons.[3]  Like Horgan, I take phenomenology seriously.  However, ‘taking it seriously’ may involve showing how phenomenal state-types could be dissociated from neurophysiological state-types.[4]  My critique will show exactly this. I will show also how the identity relation in Horgan’s view for phenomenal states presupposes what a “Unity of Science” view is and that this is a flawed outlook.  Lastly, I will show that to maintain a necessary identity between phenomenal state-types and neurophysiological state-types ignores the function which, presumably, the cognitive sciences and the philosophy of mind are charged with: finding useful ways to explain what we think of as ‘mental states.’ 

Horgan’s view begins with the premise that the phenomenology of a state cannot be argued for.  Instead, it has to do with how one looks at one’s mental life.[5]  That is, acknowledging the presence of phenomenological state R is yielded only when the individual of that state recognizes certain aspects of a mental state that can only be accessed in the first-person. The mental state may have a function, for instance, but the subjective ‘feel’ of the state, the idea that there is a way it is like to experience the state, is what Horgan means by the claim that he takes phenomenal states seriously. Thus, it is something akin to being self-evident, but it can only be realized as something self-evident when one has the proper epistemic standpoint of his first-person mental life.

Horgan’s synthesis of functionalism and type-type psychophysical identity is motivated by his critique of David Lewis’ functionalist response to the notion of inverted spectrum. Lewis’ version of functionalism maintains that names for mental states are non-rigid designators, that is, the names we use to ascribe certain mental states (e.g. the state of being in pain) have different referents across different possible worlds.  Horgan maintains this view with respect to non-phenomenal states, but argues for rigid designation for the names we use to describe phenomenal states.  Given Kripkean possible world semantics, if a name is a rigid designator, then its referent is the same in all possible worlds.  Thus if a proposition expresses a rigid designator truly, it does so necessarily.  Like the concept H20, the concept (expressed in a name) of a phenomenal state (e.g. qualia) has the same referent in all possible worlds. In contrast, given that names for non-phenomenal states are non-rigid, then any true proposition expressing a non-phenomenal state is merely contingent, since the referent of a given name of a non-phenomenal state could be different across possible worlds.  Non-phenomenal states, insofar as they are contingent and non-rigid, are characterized by contingent functional conditions. 

Horgan develops two versions of his view in accordance to the two types of functionalism he acknowledges, although the two views differ only with respect to the way non-phenomenal states are treated.  I will briefly go over the two versions of Horgan’s theory, but it is important to acknowledge that the heart of the matter for Horgan is not the way non-phenomenal states are functionally individuated, but rather the claim that phenomenal states (in both versions) “are involved in necessary psychophysical identities.” (Horgan, 460)  That said, my critique will attack the way phenomenal states are handled, and not either of the ways that non-phenomenal states incorporate one of the two kinds of functionalism.[6]

First-order functionalism “is a thoroughgoing type-type identity theory.” (Horgan, 460) This is the version of functionalism which Lewis invokes to explain the sense in which the madman and the normal human share the same state when they are in pain. Under this view, the sense in which both are in the same pain-state is possible because to be in pain is to be in the neurological state associated with pain. Horgan formally characterizes first-order functionalism as the doctrine that “each mental state-type name is definable in terms of a specific causal role, and that the name denotes whatever first-order state-type happens to occupy that role.” (Horgan, 454)  Thus for Horgan’s view, “non-phenomenal state types are identified with first-order physico-chemical state types.” (Horgan, 460) The identity relation is contingent though (as in Lewis’ view), so the concept of pain can be multiply realized in creatures that share the same neurological states but differ in how the state is normally expressed.  

Second-order functionalism is the doctrine that “each mental state-type name is definable in terms of a specific causal role, and that the name denotes a second-order state-type of the form ‘being in some first-order state-type with so-and-so typical causal role.’” This is the more typical form of functionalism, and the one most often used to debunk the classical form of the identity theory as originally argued by Stace and others.  It holds that two mental states are identical if the expression of those states occupies the same causal role in the respective systems.  The systems themselves may be materially distinct, but the way mental states are identified is via functional analysis, not microanalysis.  In this view, two different species, with different material cognitive systems, can occupy the same mental state.

Horgan’s view in total comes to the view that p: ‘phenomenal state-types are identified with the neurophysiological state-types they are expressed in (type-type identity)’ and [q: ‘first-order functionalism for non-phenomenal states’ or r: ‘second-order functionalism for non-phenomenal states.’  Note too that for any given mental state, Horgan’s view may characterize that state both according to its functional sense and according to its phenomenal sense.  That is, Horgan’s view does not require that any given mental state be one kind of state or another.  In a sense, Horgan’s view allows for a plurality of mental states.  The mental state pain may be characterized according to its causal role (second-order functionalism) and/or according to its phenomenal role (the way it is like to feel pain and the neurological state which expresses that pain).  I will now turn to a critique of p: ‘phenomenal state-types are identified with the neurophysiological state-types they are expressed in.’

Horgan’s view, with respect to phenomenal states, is somewhat motivated by what Putnam appropriately characterizes as the “Unity of Science” view. Putnam generally characterizes this view as “the doctrine that the laws of such ‘higher-level’ sciences as psychology and sociology are reducible to the laws of lower-level sciences-biology, chemistry-ultimately to the laws of elementary particle physics.”[7] (Putnam, 428) With respect to psychology’s relationship to the other sciences, the view prescribes that a complete psychology could be explainable in terms of a description of the basic constituents of the types of systems psychology characteristically deals with: namely, the cognitive systems with which it is said human agents use to interact with and form knowledge of the external world.  Cognitive theories would simply be approximations (convenient short-hands) of the more fundamental physical-chemical interactions that occur in the nervous system of individuals to which the theories apply.  It is clear that something akin to or in the vicinity of the “Unity of Science” view is applicable to meaning of Horgan’s claim that “…another positive feature of partial functionalism is that it not only accommodates this intuition, but it does so while still preserving a fully naturalistic conception of a human being as physico-chemical system whose behavior is completely explainable, in principle, solely in physico-chemical terms.” (Horgan, 461-462)

But on what grounds can the naturalistic conception of a human being as a system of physical and chemical transactions explain, for instance, the phenomenal sense of ‘being in a pain?’ Horgan is quick to draw support for his view in satisfying the intuition that the functional characterization of a state fails to speak for the experiential nature of the state.  Pain has a function but being in pain has a phenomenology which neither the causal nor the teleological senses of the concept ‘pain’ (in a functionalist theory) can sufficiently describe or explain. To subsequently hold that phenomenal state-types (for instance, the state-types of an individual who describes being in pain in terms of experiencing pain) are necessarily identical to the appropriate neurophysiological state-type(s) is not a stipulation which “takes phenomenology seriously,” but imposes another boundary, another condition which further restricts what may be asserted about phenomenal states.

The reason that positing a necessary identity relation between ascriptions of phenomenal state-types and ascriptions of neurophysiological state-types is overly restrictive is because the meaning of phenomenal state-types can be explained, articulated, given sense to, via other types of ascriptions.  There may be something about the content of a phenomenal state that is intrinsic to the experience of that state, but explaining that state, which is what the cognitive sciences and the philosophy of mind (should or does) aim to do, could involve *identifying* a given phenomenal state, or the phenomenological content of that state, with some other type of ascription. 

Putnam’s critique of this sort of identity-theorizing yields a helpful analogy:

“…given the microstructure of the brain and the nervous system, one cannot deduce that capitalist production relations will exist…the laws of capitalist society cannot be deduced from the laws of physics plus the description of the human brain: they depend on boundary conditions which are accidental from the point of view of physics but essential to the description of a situation as ‘capitalism.’” (Putnam, 431) 

That is to say, from the point of view of explaining a phenomenal state, one doesn’t arrive at an explanation via deducing the microstructure of the brain that is the vehicle of that state plus the laws of physics which govern the biochemical transactions of that state.  There is something left unexplained about the content of a state when one identifies that state-type with only one sort of state-type, and let’s face it, when one identifies state-types together, one ends up presupposing the discipline(s) which can rightfully explain, talk about, form discourses on, that one sort of state.  Phenomenal states are not necessarily identical to neurophysiological states, since the phenomenological aspect of an experienced state can be explained with recourse to, for instance, the social conditions at the time the state was experienced.[8] 

Even so, aren’t there different forms of pain?  Can one really maintain that the way pain is felt in a given situation G is identical to the way pain is felt in a given situation F?  Consider “being in abdominal pain” versus “being in emotional pain.” Ok, so Horgan may point: but look, those two situations differ with respect to the neurophysiological state-types involved.  So my theory is right.”  Yes, that may be true.  Is it necessarily true?  Of course not, neuroscientists have shown that different sorts of states occupy activate different neurological systems across different individuals.  Horgan would respond: “yes, and that’s why the phenomenal state of person S is intrinsic to that person!” I respond: Ok, but S may experience two pains simultaneously, let’s say; a pain in her foot and the emotional pain of losing a loved one.  Now let’s say that the brain scan shows her amygdala lights up and her somatosensory cortex lights up when undergoing this dual pain state-type.  A few minutes later, the brain scan shows that her amygdala is lighting up and her somatosensory cortex is also lighting up.  When asked what her experience at exactly the time the brain scan was taken, S responds: “my phenomenal state was the state of being in pain over the loss of a loved one.” This shows that there is dissociation between the activation of neurological areas and the phenomenal states which Horgan posits they are necessarily identical to. 

Given the sorts of problems I have articulated in critiquing how Horgan’s view handles phenomenal states, it seems clear that the philosophy of mind should discard the Unity of Science view, and grant that different disciplines occupy different roles with respect to explaining the phenomenal sense of one’s mental state.  Naturalism, if it means the reduction of psychological states to neurophysiological states, should be discarded in terms of a more context-friendly way of understanding the meaning of phenomenal states.  Like Wittgenstein, I propose that the meaning of something is the same as explaining it.

[1] Horgan, Terence.  Functionalism, Qualia, and Inverted Spectrum. Philosophy and Phenomenological Research. Vol. 44, No. 4 (June 1984), pp 453-469

[2] Hilary Putnam coins the term in the article Reductionism and the Nature of Psychology.  Words and Life, Harvard University Press, Cambridge, Mass.: 1994.  I explain what it means at the beginning of my criticism.

[3] Note that while I oppose the explanatory appropriateness of psychophysical identities for phenomenal states, it does not follow that such commits me to oppose the second pillar of Horgan’s view, namely functionalism for non-phenomenal states. Similarly, my opposition to the explanatory appropriateness of psychophysical identities for phenomenal states does not commit me to doubting phenomenal states in general.

[4] If one takes a subject seriously, he ought to be aware of the ways in which the matter or content of that subject could be explained. 

[5] Horgan describes how the acknowledgement of the presence of phenomenal states can’t be argued for: “…I take the intrinsic, non-relative nature of qualia to be a self-evident fact, a fact that unavoidably impresses itself upon most of us who actually experience these states…the point is…impossible to argue for…because it depends upon an individual’s first-person perspective towards his own mental life.” (Horgan, 459)

[6] Horgan maintains that the two views are equally plausible.  Thus, a critique of his view must attack the way phenomenal states are handled, and not the way either of the two ways of incorporating functionalism works (in either of the two views.)  Finally, my critique is a critique of what the two variations have in common (that phenomenal states are necessarily identified with physico-chemical states of the brain).

[7] Hilary Putnam.  Reductionism and the Nature of Psychology.  Words and Life, Harvard University Press, Cambridge, Mass.: 1994.

[8] For instance, I think it is quite clear that the phenomenal sense of ‘being aware of S’s financial problems’ is different for an S that experiences Russia’s hyperinflation in the early 1900’s as compared to the phenomenal sense of the same ascription for an S that experiences the US’s Great Depression of the 1920’s.  Recourse to the neurological states of that individual (assuming time-travel is possible) would no more point to or explain the meaning of the content of those states than it would to identify those states with the probability that it will rain according to the top officials at the time those states are experienced!


One thought on “A Critique of Horgan’s Identity for Phenomenal State-types and Neurophysiological State-types

  1. FYI this was a paper I wrote for a course I took called “20th century analytic philosophy”.

    As it is usually the case when reviewing old papers, I can definitely find things wrong with it and who knows to what extent I would stand by my arguments now. 🙂

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s